Computational Approaches in Theranostics: Mining and Predicting Cancer Data

Abstract

The ability to understand the complexity of cancer-related data has been prompted by the applications of (1) computer and data sciences, including data mining, predictive analytics, machine learning, and artificial intelligence, and (2) advances in imaging technology and probe development. Computational modelling and simulation are systematic and cost-effective tools able to identify important temporal/spatial patterns (and relationships), characterize distinct molecular features of cancer states, and address other relevant aspects, including tumor detection and heterogeneity, progression and metastasis, and drug resistance. These approaches have provided invaluable insights for improving the experimental design of therapeutic delivery systems and for increasing the translational value of the results obtained from early and preclinical studies. The big question is: Could cancer theranostics be determined and controlled in silico? This review describes the recent progress in the development of computational models and methods used to facilitate research on the molecular basis of cancer and on the respective diagnosis and optimized treatment, with particular emphasis on the design and optimization of theranostic systems. The current role of computational approaches is providing innovative, incremental, and complementary data-driven solutions for the prediction, simplification, and characterization of cancer and intrinsic mechanisms, and to promote new data-intensive, accurate diagnostics and therapeutics.

Keywords: cancer; imaging; in silico models; modeling; nanotherapeutics; simulation; theranostics.

Figure 1

Schematic illustration of the refinement process involving several integrative approaches and powered by new income and technologies. Adapted from [48]. Copyright Taylor & Francis, 2016.

Figure 2

Schematic representation of the iterative optimization workflow, reflecting the data integration/analysis and the development/optimization processes of the computational model. Reprinted from [65] under a CC BY 4.0 License. Copyright Ogilvie, L.A., Kovachev, A., Wierling, C., Lange, B.M.H. and Lehrach, H., 2017.

Figure 3

Representative 3D snapshots of the tumor response over time (t = 40, 80, 120, 200 h) and considering three drug-treatment scenarios: doxorubicin alone (Dox), sunitinib alone, and treatment with combined drugs (Dox + sunitinib). Reprinted from [111] under a CC BY 2.0 License. Copyright Wang, J., Zhang, L., Jing, C., Ye, G., Wu, H., Miao, H., Wu, Y. and Zhou, X., 2013.

Figure 4

A schematic representation of the two-dimensional projection of cellular landscape of tumor microenvironment. (a) Data from The Cancer Genome Atlas (TCGA) were analyzed and represented in a two-dimensional tumor microenvironment landscape map resorting to t-distributed stochastic neighborhood embedding (t-SNE) based on the scores reflecting the enrichment of cell types. (b) Gradient color representing the tumor metabolism index (TMI) of each sample, showing that hypometabolic samples are placed within populations with low values. (c) Circular dendrogram for the groups identified by hierarchical cluster analysis. (d) Two-dimensional map of the tumor microenvironment landscape reflecting the similarity between samples. (e) Selected clusters displaying significantly different tumor metabolic index (TMI). (f) Two-dimensional map of the tumor microenvironment landscape with the ImmuneScore information. (g) ImmuneScores were significantly different between the clusters. (h) A composed view of the results including the heatmap showing the scores of the enrichment of each immune cell type, the tumor metabolism index, the groups identified based on immune cellular heterogeneity, lung adenocarcinoma subtypes, and ImmuneScore. Reprinted from [113] under a CC BY-NC 4.0 license. Copyright Ivyspring International Publisher, 2018.

Figure 5

(a) Scheme of the modelling workflow, (b) Simulation dynamics results of the tumor growth in a virtual mouse after treatment with OT-1 T-cells and anti-CD137 monoclonal antibody, obtained at days 7, 41, 16, 20, 24, and 28 (from upper-left to lower-right). The effect of anti-CD137 in tumor infiltration is confirmed by the increased TC cell cytotoxicity and by the chemotaxis gradients. (c,d) Model prediction results obtained from the in silico study suggesting that the combined therapy with OT-1 T-cells and anti-CD137 monoclonal antibody) was not efficient in eliminating murine B16OVA melanoma with no expression of CD137 on endothelium (c). When comparing with wild-type in silico mice, the CD8 T cell infiltration was drastically reduced and delayed in time (d). Adapted from [116] under a CC BY 4.0 License. Copyright Pappalardo, F., Forero, I.M., Pennisi, M., Palazon, A., Melero, I. and Motta S., 2011.

Figure 6

Membrane interaction and internalization of the polyarabic-coated magnetite nanoparticle, MAG-ARA. (a) Initial configuration of MAR-ARA in vacuum. (b) Final configuration of MAR-ARA in contact with a dipalmitoyl phosphatidylcholine (DPPC) bilayer. (c) Example of the interactions of l-arabinose and d-galactose with DPPC. (d) Confocal fluorescence image and a bright filed image from human breast cancer cells incubated with MAG-ARA, left, and control nanoparticles, right. Reprinted from [44] under a CC BY 4.0 license. Copyright Patitsa, M., Karathanou, K., Kanaki, Z., Tzioga, L., Pippa, N., Demetzos, C., Verganelakis, D.A., Cournia, Z. and Klinakis, A., 2017.

Figure 7

(A) UV−vis absorption spectra of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs at different (AEMA):(HAuCl₄) molar ratios. (B) Transmission Electron Microscopy (TEM) image and (C) XRD plot of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs at (AEMA):(HAuCl₄) molar ratio of 5. (D) Density profiles of different beads and (E) cross-section views of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs at different (AEMA):(HAuCl₄) molar ratios. Reprinted with permission from [139]. Copyright American Chemical Society, 2017.

Figure 8

(A) Loading content, entrapment efficiency and sizes of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs/DOX as a function of m_DOX/m_micelle. (B) Density profiles of different beads and (C) cross-section views of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs/DOX micelles at different weight ratios. (D) In vitro drug release profiles of β-CD-(PCL-PAEMA-PPEGMA)21/AuNPs/DOX at different pH levels. Reprinted with permission from [139]. Copyright American Chemical Society, 2017.

Figure 9

(a,b) Illustrative description of the strategy used for detecting liver metastatic colorectal tumors using halofluorochromic polymer nanoassemblies (PNAs) and computational simulations. (c,d) Simulation results reflecting the distribution of the nanoparticles in metastatic tumors at (c) 1.8 h and (d) 12 h, after systemic injection of the nanoparticles for three different degrees of vascularization. Red refers to viable tumor tissue, blue reflects enclosing hypoxic tissue, and brown reflects necrotic regions. The rectangular grid represents the capillary network, displaying irregular sprouts, which simulate the angiogenesis effect and the growth of blood vessel. Reprinted with permission from [147]. Copyright Springer Nature, 2017.

Figure 10

(a–c) General procedure of the spectral Fiedler field (SFF) method which includes (a) a raw image in gray scale, (b) the matrix representation of (a), and (c) the resulting contrast matrix. Using the matrix from of each gray scale image, the method is applied to a small region centered at every single pixel, generating the contrast matrix, which is then processed using a low pass filter for removing noise. The final matrix represents the SFF map. (d) Hematoxylin and eosin image of colon tumor. (e) Pattern obtained using CD31 endothelial cell marker. (f) Fluorescence of the drug in tumors. (g) SFF map of echogenic liposomes enhanced ultrasound contrast in tumor areas displaying a positive response for CD31 endothelial cell marker. (h) Relative drug diffusion considering the positive regions in SFF map. (i) Normalized distribution of doxorubicin combined with doxorubicin fluorescence image. Purple shades refer to doxorubicin fluorescence while blue-red corresponds to the spectral shades of doxorubicin distribution. Reprinted from [175] under a CC BY 4.0 License. Copyright Liu, C., Kapoor, A., VanOsdol, J., Ektate, K., Kong, Z., and Ranjan, A., 2018.